The problem with the XBS170 subcircuit is that the values of w and l,
which are both 1, are too large. If you change them to smaller values,
hspice runs without aborting. I just changed both w and l to 1u instead
and my simulations matched up almost perfectly with my measured values.
Hope this helps.
Alex Sun